CN113672489B - Resource performance level determination method and equipment for super computer - Google Patents
Resource performance level determination method and equipment for super computer Download PDFInfo
- Publication number
- CN113672489B CN113672489B CN202111237139.4A CN202111237139A CN113672489B CN 113672489 B CN113672489 B CN 113672489B CN 202111237139 A CN202111237139 A CN 202111237139A CN 113672489 B CN113672489 B CN 113672489B
- Authority
- CN
- China
- Prior art keywords
- state value
- memory
- target
- supercomputer
- cpu
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3466—Performance evaluation by tracing or monitoring
- G06F11/3476—Data logging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3055—Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Computer Hardware Design (AREA)
- Debugging And Monitoring (AREA)
Abstract
The application provides a resource performance level determination method and device of a super computer, computer equipment and a storage medium, relates to the technical field of super computing, and is used for improving the utilization rate of resources of the super computer. The method mainly comprises the following steps: acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer; adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value; and determining the resource performance level of the target super computer according to the system state value and the application state value.
Description
Technical Field
The present application relates to the field of supercomputer technologies, and in particular, to a supercomputer evaluation method and apparatus, a computer device, and a storage medium.
Background
The supercomputer is a computer which is formed by combining a plurality of computing nodes and can perform large-scale computation or data processing in parallel, is also called as a parallel computer, is the computer with the strongest function, the fastest operation and the largest storage capacity, is mainly used for the national high-tech field and the advanced technical research, and is an important embodiment of the national science and technology development level and the comprehensive national force.
At present, more and more super computers are provided with more and more architecture and technical routes, super computing centers in various places, such as bamboo shoots in spring after raining, are in endless, and in the face of such many super computers, users face difficulty in selection, and it is unclear which is better to select.
Disclosure of Invention
The embodiment of the application provides a resource performance level determination method and device for a supercomputer, computer equipment and a storage medium, which are used for improving the utilization rate of supercomputer resources.
The embodiment of the invention provides a resource performance level determination method of a super computer, which comprises the following steps:
acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer;
adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value;
and determining the resource performance level of the target super computer according to the system state value and the application state value.
The embodiment of the invention provides a resource performance level determining device of a super computer, which comprises:
the system comprises an acquisition module, a storage module and a control module, wherein the acquisition module is used for acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target super computer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target super computer;
the computing module is used for adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value;
and the determining module is used for determining the resource performance level of the target super computer according to the system state value and the application state value.
A computer device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, said processor implementing the resource performance level determination method of the above-mentioned supercomputer when executing said computer program.
A computer-readable storage medium, in which a computer program is stored, which computer program, when being executed by a processor, carries out the method of resource performance level determination of a supercomputer as described above.
The invention provides a resource performance grade determination method, a device, computer equipment and a storage medium of a supercomputer.A Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer are firstly obtained; then adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value; and finally, determining the resource performance level of the target supercomputer according to the system state value and the application state value. Therefore, the resource performance owned by the super computer can be calculated through the method, the resource performance level corresponding to the super computer can be determined, and the resource utilization rate of the super computer is improved according to the resource new energy level of the super computer.
Drawings
FIG. 1 is a flow chart of a resource performance level determination method for a supercomputer according to the present application;
fig. 2 is a schematic diagram illustrating a memory flow required by an acquisition program according to the present application;
fig. 3 is a schematic structural diagram of a resource performance level determination apparatus of a supercomputer according to the present application.
Detailed Description
In order to better understand the technical solutions described above, the technical solutions of the embodiments of the present application are described in detail below with reference to the drawings and the specific embodiments, and it should be understood that the specific features of the embodiments and the embodiments of the present application are detailed descriptions of the technical solutions of the embodiments of the present application, and are not limitations of the technical solutions of the present application, and the technical features of the embodiments and the embodiments of the present application may be combined with each other without conflict.
Referring to FIG. 1, a method for determining the performance level of a resource of a supercomputer according to the present invention is shown. The method specifically comprises the steps of S10-S30:
step S10, obtaining a CPU state value, a memory state value, a network state value, and an input/output I/O state value of the target supercomputer, and an operating time state value, a CPU utilization rate, and a memory utilization rate required for the target application to operate under the target supercomputer.
Wherein, the target super computer is a computer needing to determine the performance level of the resource. For example, the target supercomputer may be a Tianhe supercomputer Tianhe 1A (TH 1A), a Tianhe model III prototype (TH 3E), a self-built HPC cluster (HPC) of the national supercomputing Tianjin center, and the like, and the embodiment is not particularly limited.
In this embodiment, the CPU state value, the memory state value, the network state value, and the input/output I/O state value are respectively used to represent the CPU performance, the memory performance, the network performance, and the I/O performance of the target supercomputer. Certainly, fine-grained state values of the target supercomputer, such as power consumption test, high-temperature test, pressure test, noise test and the like, can also be obtained, and then calculation is performed according to the state values of all dimensions of the target supercomputer to obtain the system state value of the target supercomputer.
The running time state value is used for evaluating the time occupied by the target application running in the target super computer, the CPU utilization rate is used for evaluating the CPU resource occupied by the target application running in the target super computer, and the memory utilization rate is used for evaluating the memory resource occupied by the target application running in the target super computer.
In an embodiment of the present invention, the step S10 of obtaining the CPU state value, the memory state value, the network state value, and the I/O state value of the target supercomputer includes:
step S101, acquiring first intermediate parameters and second intermediate parameters corresponding to different evaluation types of the target supercomputer.
The evaluation type comprises a CPU, a memory, a network and an I/O. The first intermediate parameter is an intermediate parameter with higher numerical value and better performance, and for example, the first intermediate parameter may specifically be a CPU floating point operation speed, a memory bandwidth, a network bandwidth, a CPU utilization rate, a memory utilization rate, and the like; the second intermediate parameter is an intermediate parameter with a smaller numerical value and better performance, for example, the second intermediate parameter may specifically be running time, network delay, and the like, and the embodiment of the present invention is not specifically limited.
Step S102, calculating the state value of the first intermediate parameter by the following formula:
step S103, calculating the state value of the second intermediate parameter by the following formula:
and step S104, calculating the sum of the state value of the first intermediate parameter and the state value of the second intermediate parameter belonging to the same evaluation type to obtain a CPU state value, a memory state value, a network state value and an I/O state value.
Wherein the content of the first and second substances,for the target supercomputer atThe result of the evaluation of the aspect,in order for the target super-computer to be,a first intermediate parameter or a second intermediate parameter respectively corresponding to the evaluation type CPU, the memory, the network and the I/O,to representHigher coefficients, which means that the intermediate parameters are more important, generally coefficients in terms of cpu are most important, followed by network, memory and I/O,,is a fixed value set in advance.
for presetting supercomputersIn thatAspect evaluation result, presetting supercomputerAnd the basic supercomputer is used for measuring the evaluation result of the target supercomputer. In the present embodiment, it is preferred that,benckmarkcomparisons need to be made at equal core numbers.
It should be noted that, in the present embodiment, each performance intermediate parameter of the target supercomputer may be obtained by the system evaluation tool. For example, the embodiment is not specifically limited to this, and the method includes obtaining a memory bandwidth test result through a tool Stream, obtaining a memory pressure test result through a tool memtest, obtaining a metadata operation test result through a tool Mdtest, obtaining a file system test result through a tool IOR, obtaining an IO performance test result through a tool IOzone, obtaining a network performance test result through a tool Osu, obtaining a CPU performance test result through a tool linepack, and the like.
In another embodiment provided by the present invention, the step S10 obtains the runtime state value, the CPU utilization, and the memory utilization required by the target application to run under the target supercomputer, and includes:
and S101, acquiring a third intermediate parameter and a fourth intermediate parameter respectively corresponding to different evaluation types of the target application under the target super computer.
The evaluation type comprises an operation time state value, a CPU utilization rate and a memory utilization rate; the third intermediate parameter is an intermediate parameter with higher numerical value and better performance, and for example, the third intermediate parameter may specifically be a CPU floating point operation speed, a memory bandwidth, a network bandwidth, a CPU utilization rate, a memory utilization rate, and the like; the fourth intermediate parameter is an intermediate parameter with a smaller numerical value and better performance, for example, the fourth intermediate parameter may specifically be running time, network delay, and the like, and the embodiment of the present invention is not specifically limited.
Step S102, calculating the state value of the third intermediate parameter by the following formula:
step S103, calculating the state value of the fourth intermediate parameter by the following formula:
step S104, calculating the sum of the state value of the third intermediate parameter and the state value of the fourth intermediate parameter belonging to the same evaluation type to obtain an operating time state value, a CPU utilization rate and a memory utilization rate;
wherein the content of the first and second substances,running on the target supercomputer for the target applicationThe result of the evaluation of the aspect,for the purpose of the said target application(s),in order for the target super-computer to be,running time status value, CPU utilization, and memory utilization score for evaluation typeRespectively corresponding third intermediate parameter or fourth intermediate parameter,to representHigher coefficient, indicating that the intermediate parameter is more important,,the value is a fixed value set in advance, and is usually set to 10.
For the target applicationAt the target supercomputerOn the upper partEvaluation results of aspects;
for the target applicationOn a preset super computerOn the upper partAnd (4) evaluating results of aspects. Preset super computerAnd the basic supercomputer is used for measuring the evaluation result of the target supercomputer. In the present embodiment, it is preferred that,benckmarkcomparisons need to be made at equal core numbers.
Step S20, add the CPU state value, the memory state value, the network state value, and the I/O state value to obtain a system state value, and add the operating time state value, the CPU utilization rate, and the memory utilization rate to obtain an application state value.
In particular, the method is characterized by the formulaCalculating a CPU state value, a memory state value, a network state value and an I/O state value, and then adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state valueBy the above formulaCalculating the running time state value, the CPU utilization rate and the memory utilization rate, and then adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value 。
And step S30, determining the resource performance level of the target super computer according to the system state value and the application state value.
In particular, by the formulaA resource performance level of the target supercomputer is determined. The resource performance level may be a specific state value, or after obtaining a specific state value, the resource performance level may be determined according to the state value. More specifically, the present embodiment may divide the supercomputer into N according to actual requirementsThe resource performance levels, such as the first level, the second level, and the third level, may be represented by the first level as the optimal resource performance of the supercomputer, the second level as the good resource performance of the supercomputer, and the third level as the poor resource performance level of the supercomputer, which is not specifically limited in this embodiment.
The invention provides a resource performance grade determination method of a supercomputer, which comprises the steps of firstly obtaining a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and obtaining an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer; then adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value; and finally, determining the resource performance level of the target supercomputer according to the system state value and the application state value. Therefore, the resource performance owned by the super computer can be calculated through the method, the resource performance level corresponding to the super computer can be determined, and the resource utilization rate of the super computer is improved according to the resource new energy level of the super computer.
In one embodiment provided by the present invention, the method may further include:
step S201Programs of different calculation sizesAre respectively atDifferent supercomputersProvided withA plurality of different computing resourcesAnd running to obtain the running time state value, the CPU utilization rate and the memory utilization rate under the corresponding relation among the program, the supercomputer and the computing resources.
And the memory of the node corresponding to the computing resource is larger than the memory required for running the corresponding program.
For example, the user-provided willPrograms of different calculation sizesAre respectively atDifferent supercomputersProvided withA plurality of different computing resourcesThe operation is carried out. Wherein computing resourcesTaking a computing node providing high-performance computing power as a unit, the number of distributed nodes is required to meet the condition that the total memory of the computing node is larger than the memory required by running a corresponding programThen, the corresponding program running time is calculated by using a monitoring tool integrated with the platform
And application run key features: CPU utilization
And application run key features: memory utilization
Wherein the content of the first and second substances,for the program run time(s),CPU resource utilization (%) occupied for the application,calculating the scale for the memory resource utilization rate (%) occupied by the application programAre all according to the factThe actual test scale is scaled to the scaled test scale.
Step S202, determining an operation time state value fitting function, a CPU utilization ratio fitting function and a memory utilization ratio fitting function according to the operation time state value, the CPU utilization ratio and the memory utilization ratio under the corresponding relation of the program, the supercomputer and the computing resources.
Calculating the program running time by the running time state value, the CPU utilization rate and the memory utilization rate under the corresponding relation among the program, the supercomputer and the calculation resources obtained in the step S201Program CPU utilizationAnd memory utilizationEstimate the example sizeAnd using computing resourcesIn relation to (2)。
At run timeFor example, assume a nonlinear curve function of the formAnd satisfyWherein, in the step (A),,then, the result is substituted into an lsqcurvefit function of Matlab, and the least square method is used for estimation, so that a fitting function can be calculated. By the same token, the method can obtain. It should be noted that the form of the nonlinear curve function may be modified according to actual situations, and this embodiment is not particularly limited thereto.
Step S203, calculating the running of the target application on the target super computer through the running time state value fitting function, the CPU utilization ratio fitting function and the memory utilization ratio fitting functionAnd (4) evaluating results of aspects.
In another embodiment provided by the present invention, as shown in fig. 2, the obtaining the memory required by the program includes:
step S301 of acquiring the programAnd respectively evaluating the required memories of the operation according to the corresponding histories.
Step S302, determining the memory required by the historical evaluation operation as an initial memory。
Step S303, passing formula= And calculating the memory required by running the corresponding program.
Wherein the content of the first and second substances,and calculating the node memory on the cluster.
Furthermore, after the memory required for running the corresponding program, the program needs to be monitored and run in real time by calculationOccupied memory value; if the memory value exceeds a preset value, passing a formulaRe-determining the memory required by running the corresponding program; wherein the content of the first and second substances,,。
it should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
In an embodiment, a resource performance level determining apparatus of a super computer is provided, where the resource performance level determining apparatus of the super computer corresponds to the resource performance level determining method of the super computer in the above embodiments one to one. As shown in fig. 3, the resource performance level determination apparatus of the super computer has the following functional modules:
an obtaining module 10, configured to obtain a CPU state value, a memory state value, a network state value, and an input/output I/O state value of a central processing unit of a target supercomputer, and an operating time state value, a CPU utilization rate, and a memory utilization rate that are required by a target application to operate under the target supercomputer;
a calculating module 20, configured to add the CPU state value, the memory state value, the network state value, and the I/O state value to obtain a system state value, and add the operating time state value, the CPU utilization, and the memory utilization to obtain an application state value;
and the determining module 30 is configured to determine the resource performance level of the target supercomputer according to the system state value and the application state value.
In an embodiment that may be provided by the present invention, the obtaining module 10 is specifically configured to:
acquiring a first intermediate parameter and a second intermediate parameter which respectively correspond to different evaluation types of the target supercomputer, wherein the evaluation types comprise a CPU (Central processing Unit), a memory, a network and an I/O (input/output);
calculating a state value of the first intermediate parameter by:
calculating a state value of the second intermediate parameter by the following formula:
obtaining a CPU state value, a memory state value, a network state value and an I/O state value by calculating the sum of the state value of the first intermediate parameter and the state value of the second intermediate parameter belonging to the same evaluation type;
wherein the content of the first and second substances,for the target supercomputer atThe result of the evaluation of the aspect,in order for the target super-computer to be,a first intermediate parameter or a second intermediate parameter respectively corresponding to the evaluation type CPU, the memory, the network and the I/O,to representThe significant coefficient of (a) of (b),,is a preset fixed value;
In another embodiment that may be provided by the present invention, the obtaining module 10 is specifically configured to:
acquiring a third intermediate parameter and a fourth intermediate parameter respectively corresponding to different evaluation types of a target application under the target super computer, wherein the evaluation types comprise an operation time state value, a CPU utilization rate and a memory utilization rate;
calculating a state value of the third intermediate parameter by:
calculating a state value of the fourth intermediate parameter by:
obtaining an operating time state value, a CPU utilization rate and a memory utilization rate by calculating a sum of the state value of the third intermediate parameter and the state value of the fourth intermediate parameter belonging to the same evaluation type;
wherein the content of the first and second substances,running on the target supercomputer for the target applicationThe result of the evaluation of the aspect,for the purpose of the said target application(s),in order for the target super-computer to be,a third intermediate parameter or a fourth intermediate parameter respectively corresponding to the evaluation type running time state value, the CPU utilization rate and the memory utilization rate,to representThe significant coefficient of (a) of (b),,is a preset fixed value;
for the target applicationAt the target supercomputerOn the upper partEvaluation results of aspects;
for the target applicationOn a preset super computerOn the upper partAnd (4) evaluating results of aspects.
Further, the obtaining module 10 is also used for obtainingPrograms of different calculation sizesAre respectively atDifferent supercomputersProvided withA plurality of different computing resourcesRunning to obtain a running time state value, a CPU utilization rate and a memory utilization rate under the corresponding relation among the program, the supercomputer and the computing resources; the memory of the node corresponding to the computing resource is larger than the memory required by running the corresponding program;
further, the determining module 30 is further configured to determine an operation time state value fitting function, a CPU utilization ratio fitting function, and a memory utilization ratio fitting function according to the operation time state value, the CPU utilization ratio, and the memory utilization ratio in the corresponding relationship among the program, the supercomputer, and the computing resources;
further, the calculating module 20 is further configured to calculate the target application running on the target supercomputer according to the running time state value fitting function, the CPU utilization ratio fitting function, and the memory utilization ratio fitting functionAnd (4) evaluating results of aspects.
Further, the obtaining module 10 is further configured to obtain the programRespectively corresponding memories required by historical evaluation operation;
further, the determining module 30 is further configured to determine the memory required by the historical evaluation operation as an initial memory;
Further, the calculating module 20 is further configured to calculate the formula= Calculating a memory required for running a corresponding program; wherein the content of the first and second substances,and calculating the node memory on the cluster.
Further, the computing module 20 is further configured to monitor and operate the program in real timeOccupied memory value; if the memory value exceeds a preset value, passing a formulaRe-determining the memory required by running the corresponding program; wherein the content of the first and second substances,,。
for the specific limitation of the resource performance level determination apparatus of the super computer, reference may be made to the above limitation of the resource performance level determination method of the super computer, and details are not described here. The various modules in the above-described apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:
acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer;
adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value;
and determining the resource performance level of the target super computer according to the system state value and the application state value.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer;
adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value;
and determining the resource performance level of the target super computer according to the system state value and the application state value.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-mentioned functions.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.
Claims (9)
1. A method for resource performance level determination for a supercomputer, the method comprising:
acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target supercomputer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target supercomputer;
adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value;
determining the resource performance level of the target super computer according to the system state value and the application state value;
will be provided withPrograms of different calculation sizesAre respectively atDifferent supercomputersProvided withA plurality of different computing resourcesRunning to obtain a running time state value, a CPU utilization rate and a memory utilization rate under the corresponding relation among the program, the supercomputer and the computing resources; the memory of the node corresponding to the computing resource is larger than the memory required by running the corresponding program;
determining an operation time state value fitting function, a CPU utilization ratio fitting function and a memory utilization ratio fitting function according to the operation time state value, the CPU utilization ratio and the memory utilization ratio under the corresponding relation among the program, the supercomputer and the computing resources;
2. The method of claim 1, wherein obtaining the CPU state value, the memory state value, the network state value, and the I/O state value of the target supercomputer comprises:
acquiring a first intermediate parameter and a second intermediate parameter which respectively correspond to different evaluation types of the target supercomputer, wherein the evaluation types comprise a CPU (Central processing Unit), a memory, a network and an I/O (input/output);
calculating a state value of the first intermediate parameter by:
calculating a state value of the second intermediate parameter by the following formula:
obtaining a CPU state value, a memory state value, a network state value and an I/O state value by calculating the sum of the state value of the first intermediate parameter and the state value of the second intermediate parameter belonging to the same evaluation type;
wherein the content of the first and second substances,for the target supercomputer atThe result of the evaluation of the aspect,in order for the target super-computer to be,a first intermediate parameter or a second intermediate parameter respectively corresponding to the evaluation type CPU, the memory, the network and the I/O,to representThe significant coefficient of (a) of (b),,is a preset fixed value;
3. The method of claim 1, wherein obtaining the run-time state value, CPU utilization, and memory utilization required by the target application to run under the target supercomputer comprises
Acquiring a third intermediate parameter and a fourth intermediate parameter respectively corresponding to different evaluation types of a target application under the target super computer, wherein the evaluation types comprise an operation time state value, a CPU utilization rate and a memory utilization rate;
calculating a state value of the third intermediate parameter by:
calculating a state value of the fourth intermediate parameter by:
obtaining an operating time state value, a CPU utilization rate and a memory utilization rate by calculating a sum of the state value of the third intermediate parameter and the state value of the fourth intermediate parameter belonging to the same evaluation type;
wherein the content of the first and second substances,running on the target supercomputer for the target applicationThe result of the evaluation of the aspect,for the purpose of the said target application(s),in order for the target super-computer to be,a third intermediate parameter or a fourth intermediate parameter respectively corresponding to the evaluation type running time state value, the CPU utilization rate and the memory utilization rate,to representThe significant coefficient of (a) of (b),,is a preset fixed value;
for the target applicationAt the target supercomputerOn the upper partEvaluation results of aspects;
4. The method of claim 1, further comprising:
acquiring the programRespectively corresponding memories required by historical evaluation operation;
5. The method of claim 4, further comprising:
if the memory value exceeds a preset value, passing a formulaRe-determining the memory required by running the corresponding program;
6. An apparatus for determining a performance level of a resource of a supercomputer, the apparatus comprising:
the system comprises an acquisition module, a storage module and a control module, wherein the acquisition module is used for acquiring a Central Processing Unit (CPU) state value, a memory state value, a network state value and an input/output (I/O) state value of a target super computer, and an operation time state value, a CPU utilization rate and a memory utilization rate which are required by the running of a target application under the target super computer;
the computing module is used for adding the CPU state value, the memory state value, the network state value and the I/O state value to obtain a system state value, and adding the running time state value, the CPU utilization rate and the memory utilization rate to obtain an application state value;
the determining module is used for determining the resource performance level of the target super computer according to the system state value and the application state value;
an acquisition module for further processingPrograms of different calculation sizesAre respectively atDifferent supercomputersProvided withA plurality of different computing resourcesRunning to obtain a running time state value, a CPU utilization rate and a memory utilization rate under the corresponding relation among the program, the supercomputer and the computing resources; the memory of the node corresponding to the computing resource is larger than the memory required by running the corresponding program;
the determining module is further configured to determine an operating time state value fitting function, a CPU utilization ratio fitting function, and a memory utilization ratio fitting function according to the operating time state value, the CPU utilization ratio, and the memory utilization ratio in the correspondence among the program, the supercomputer, and the computing resources;
the calculation module is further configured to calculate the target application running on the target supercomputer according to the running time state value fitting function, the CPU utilization ratio fitting function, and the memory utilization ratio fitting functionAnd (4) evaluating results of aspects.
7. The apparatus of claim 6, wherein the obtaining module is specifically configured to:
acquiring a first intermediate parameter and a second intermediate parameter which respectively correspond to different evaluation types of the target supercomputer, wherein the evaluation types comprise a CPU (Central processing Unit), a memory, a network and an I/O (input/output);
calculating a state value of the first intermediate parameter by:
calculating a state value of the second intermediate parameter by the following formula:
obtaining a CPU state value, a memory state value, a network state value and an I/O state value by calculating the sum of the state value of the first intermediate parameter and the state value of the second intermediate parameter belonging to the same evaluation type;
wherein the content of the first and second substances,for the target supercomputer atThe result of the evaluation of the aspect,in order for the target super-computer to be,a first intermediate parameter or a second intermediate parameter respectively corresponding to the evaluation type CPU, the memory, the network and the I/O,to representThe significant coefficient of (a) of (b),,is a preset fixed value;
8. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the resource performance level determination method of a supercomputer according to any one of claims 1 to 5 when executing the computer program.
9. A computer-readable storage medium, in which a computer program is stored, which, when being executed by a processor, carries out a method of resource performance level determination of a supercomputer according to any one of claims 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111237139.4A CN113672489B (en) | 2021-10-25 | 2021-10-25 | Resource performance level determination method and equipment for super computer |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111237139.4A CN113672489B (en) | 2021-10-25 | 2021-10-25 | Resource performance level determination method and equipment for super computer |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113672489A CN113672489A (en) | 2021-11-19 |
CN113672489B true CN113672489B (en) | 2022-01-25 |
Family
ID=78550966
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111237139.4A Active CN113672489B (en) | 2021-10-25 | 2021-10-25 | Resource performance level determination method and equipment for super computer |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113672489B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115686865B (en) * | 2022-11-18 | 2023-07-07 | 天河超级计算淮海分中心 | Super computing node resource distribution system based on multi-scene application |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103412806A (en) * | 2013-08-12 | 2013-11-27 | 浪潮电子信息产业股份有限公司 | Evaluation method for comprehensive performance of multi-category applications used on super computer |
CN104199744A (en) * | 2014-08-29 | 2014-12-10 | 浪潮(北京)电子信息产业有限公司 | Method and device for judging performance stability of applications of super computer |
CN104252540A (en) * | 2014-09-24 | 2014-12-31 | 杭州顺网科技股份有限公司 | Game configuration push method based on computer performance evaluation |
CN105573898A (en) * | 2015-12-11 | 2016-05-11 | 中国航空工业集团公司西安航空计算技术研究所 | Automatic test and evaluation method for comprehensive performance of airborne computer |
CN105740126A (en) * | 2016-03-08 | 2016-07-06 | 西北工业大学 | Embedded system performance evaluation method based on five capabilities |
CN106250306A (en) * | 2016-08-18 | 2016-12-21 | 电子科技大学 | A kind of performance prediction method being applicable to enterprise-level O&M automatization platform |
CN108334440A (en) * | 2017-01-19 | 2018-07-27 | 阿里巴巴集团控股有限公司 | A kind of processing method and processing device, client obtaining application performance test result |
CN108509315A (en) * | 2018-03-26 | 2018-09-07 | 中国电子产品可靠性与环境试验研究所((工业和信息化部电子第五研究所)(中国赛宝实验室)) | Processor performance evaluates and tests comparative approach, device, computer equipment and storage medium |
CN111813644A (en) * | 2020-07-16 | 2020-10-23 | 中国民航信息网络股份有限公司 | System performance evaluation method and device, electronic equipment and computer readable medium |
CN111881009A (en) * | 2020-06-29 | 2020-11-03 | 苏州浪潮智能科技有限公司 | Method, device, medium and computer equipment for evaluating super computing performance |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090144744A1 (en) * | 2007-11-29 | 2009-06-04 | Gunnels John A | Performance Evaluation of Algorithmic Tasks and Dynamic Parameterization on Multi-Core Processing Systems |
CN103902443B (en) * | 2012-12-26 | 2017-04-26 | 华为技术有限公司 | Program running performance analysis method and device |
US9558094B2 (en) * | 2014-05-12 | 2017-01-31 | Palo Alto Research Center Incorporated | System and method for selecting useful smart kernels for general-purpose GPU computing |
US9971391B2 (en) * | 2015-12-23 | 2018-05-15 | Intel Corporation | Method to assess energy efficiency of HPC system operated with and without power constraints |
CN107562532B (en) * | 2017-07-13 | 2020-11-27 | 华为技术有限公司 | Method and device for predicting hardware resource utilization rate of equipment cluster |
EP3671465B1 (en) * | 2018-12-21 | 2021-08-11 | Bull SAS | Method for analysing the performance of a software application on a cluster of servers by extrapolation |
CN110673954B (en) * | 2019-09-18 | 2022-02-15 | 北京北科融智云计算科技有限公司 | Utilization rate prediction method, utilization rate prediction device, computing equipment, storage medium and program product |
-
2021
- 2021-10-25 CN CN202111237139.4A patent/CN113672489B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103412806A (en) * | 2013-08-12 | 2013-11-27 | 浪潮电子信息产业股份有限公司 | Evaluation method for comprehensive performance of multi-category applications used on super computer |
CN104199744A (en) * | 2014-08-29 | 2014-12-10 | 浪潮(北京)电子信息产业有限公司 | Method and device for judging performance stability of applications of super computer |
CN104252540A (en) * | 2014-09-24 | 2014-12-31 | 杭州顺网科技股份有限公司 | Game configuration push method based on computer performance evaluation |
CN105573898A (en) * | 2015-12-11 | 2016-05-11 | 中国航空工业集团公司西安航空计算技术研究所 | Automatic test and evaluation method for comprehensive performance of airborne computer |
CN105740126A (en) * | 2016-03-08 | 2016-07-06 | 西北工业大学 | Embedded system performance evaluation method based on five capabilities |
CN106250306A (en) * | 2016-08-18 | 2016-12-21 | 电子科技大学 | A kind of performance prediction method being applicable to enterprise-level O&M automatization platform |
CN108334440A (en) * | 2017-01-19 | 2018-07-27 | 阿里巴巴集团控股有限公司 | A kind of processing method and processing device, client obtaining application performance test result |
CN108509315A (en) * | 2018-03-26 | 2018-09-07 | 中国电子产品可靠性与环境试验研究所((工业和信息化部电子第五研究所)(中国赛宝实验室)) | Processor performance evaluates and tests comparative approach, device, computer equipment and storage medium |
CN111881009A (en) * | 2020-06-29 | 2020-11-03 | 苏州浪潮智能科技有限公司 | Method, device, medium and computer equipment for evaluating super computing performance |
CN111813644A (en) * | 2020-07-16 | 2020-10-23 | 中国民航信息网络股份有限公司 | System performance evaluation method and device, electronic equipment and computer readable medium |
Non-Patent Citations (1)
Title |
---|
"面向ARM架构的HPC***性能评测及应用移植优化";戈孜荣;《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》;20210115(第01期);论文第三章-第四章 * |
Also Published As
Publication number | Publication date |
---|---|
CN113672489A (en) | 2021-11-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10963292B2 (en) | Techniques to manage virtual classes for statistical tests | |
Khaleghzadeh et al. | A novel data-partitioning algorithm for performance optimization of data-parallel applications on heterogeneous HPC platforms | |
CN108205469B (en) | MapReduce-based resource allocation method and server | |
CN113672489B (en) | Resource performance level determination method and equipment for super computer | |
CN113296905A (en) | Scheduling method, scheduling device, electronic equipment, storage medium and software product | |
Beard et al. | Analysis of a simple approach to modeling performance for streaming data applications | |
US20150331787A1 (en) | Software verification | |
CN104731563A (en) | FFT-based large integer multiplication SSA algorithm multi-core parallel implementation method | |
Zeni et al. | Optimized implementation of the hpcg benchmark on reconfigurable hardware | |
Lewis et al. | Chaotic attractor prediction for server run-time energy consumption | |
Alonso et al. | Experimental study of six different implementations of parallel matrix multiplication on heterogeneous computational clusters of multicore processors | |
CN113158435B (en) | Complex system simulation running time prediction method and device based on ensemble learning | |
Rizvandi et al. | On modeling dependency between mapreduce configuration parameters and total execution time | |
Vikranth et al. | Topology aware task stealing for on-chip NUMA multi-core processors | |
CN104281636A (en) | Concurrent distributed processing method for mass report data | |
Bernabé et al. | Tuning basic linear algebra routines for hybrid CPU+ GPU platforms | |
CN116303219A (en) | Grid file acquisition method and device and electronic equipment | |
CN114021733B (en) | Model training optimization method, device, computer equipment and storage medium | |
CN116069603B (en) | Performance test method of application, method and device for establishing performance test model | |
CN113886057B (en) | Dynamic resource scheduling method based on analysis technology and data stream information on heterogeneous many-core | |
Hauck et al. | Qualitative properties of mathematical model for data flow | |
Muresano et al. | Combining scalability and efficiency for spmd applications on multicore clusters | |
CN105512087B (en) | Reliability evaluation method of resource-constrained multi-node computing system | |
Mudalige et al. | Predictive analysis and optimisation of pipelined wavefront computations | |
Zhang et al. | Approximate mean value analysis for multi-core systems |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |